Open AccessAssociation of peripheral blood neutrophil gelatinase‐associated lipocalin levels with bone marrow neutrophil gelatinase‐associated lipocalin levels and neutrophil count in hematologic malignancy
Author(s) -
Cho ChiHyun,
Yoon Jung,
Kim DeokSu,
Kim ShinJong,
Sung Hwa Jung,
Lee Se Ryeon
Publication year - 2019
Publication title -
journal of clinical laboratory analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.536
H-Index - 50
eISSN - 1098-2825
pISSN - 0887-8013
DOI - 10.1002/jcla.22920
Subject(s) - absolute neutrophil count , lipocalin , medicine , neutrophil gelatinase associated lipocalin , bone marrow , hematologic malignancy , hematology , complete blood count , gastroenterology , biomarker , count data , neutropenia , immunology , biology , toxicity , biochemistry , statistics , mathematics , poisson distribution
Background Although neutrophil gelatinase‐associated lipocalin (NGAL) is a biomarker for acute kidney injury, recently, high NGAL levels have been reported in hematologic malignancies. Given the mechanism underlying NGAL synthesis and secretion in neutrophilic series, it is speculated that NGAL levels are higher in bone marrow (BM) than in peripheral blood (PB). Additionally, PB NGAL levels are thought to be associated with neutrophilic parameters. We aimed to test both hypotheses in hematologic malignancies. Methods Paired BM and PB samples were collected from 41 patients undergoing BM examination for hematologic malignancies. NGAL levels were measured using immunoassays. Data on hematologic parameters were collected from medical records. Single and multiple regression analyses were performed to analyze the relationship. Results PB and BM NGAL (n = 41) levels were significantly different (163.0 ± 258.3 and 413.1 ± 616.2 ng/mL [mean ± standard deviation], respectively; P < 0.05). Simple regression analysis and multicollinearity assessment showed that BM NGAL levels, BM neutrophil%, and neutrophil count were significant predictors of PB NGAL. Two multiple regression models were developed (model 1, PB NGAL = 21.467* neutrophil count ‐ 0.785*BM neutrophil%; model 2, PB NGAL = 21.202*neutrophil count‐ 0.915*BM neutrophil% +0.10*BM NGAL). Akaike's information criterion and adjusted R 2 values showed that model 1 had higher predictive accuracy for PB NGAL. In both models, neutrophil count was the only significant predictor. Conclusion BM NGAL was significantly higher than PB NGAL in hematologic malignancy. In addition, PB NGAL could be expressed as a multiple regression model including neutrophil count and BM neutrophil%, being significantly influenced by neutrophil count.